Europe controls 10% of AI compute. A new report asks what that costs

Europe controls roughly 10% of global AI data centre capacity. The United States and China hold the rest. That single statistic, drawn from a new Deloitte scenario analysis of European technological sovereignty, captures the scale of the continent's dependency. Over the next decade, Europe will either close that gap or find its industries, governments and citizens operating on infrastructure owned and governed elsewhere.

Deloitte's Centre for the Long View has published four scenarios for European technology by 2036, built from 90 identified drivers and refined through expert panels. The exercise is deliberately extreme in its range: one scenario envisages European digital excellence generating annual revenues above €3 trillion; others imply stagnation or continued dependency. The point is not prediction but preparation. The two forces that shape all four scenarios are the same ones shaping policy today: the advance of artificial intelligence, and the fragmentation of global supply chains along geopolitical lines.

Europe's position is not ambiguous. It runs third in the AI race, behind the United States and China, and the gap is widening rather than narrowing. The continent has a limited record of commercialising its own research, and it struggles to retain the talent that produces it. Deloitte frames this as a double deficit: Europe lacks both the supply chain independence that faster-growing regions enjoy, and the technology leadership that drives growth in those same regions.

The dependency extends beyond compute. Access to raw materials, semiconductor manufacturing capacity, patents and data infrastructure are all contested in the strategic rivalry between Washington and Beijing. Europe is a demander, not a supplier, in most of these chains.

The compute question

Artificial intelligence is shifting from a specialist tool to something closer to critical infrastructure, comparable to energy networks. The comparison matters because Europe treats energy infrastructure as a sovereignty concern. It does not yet treat compute the same way. More than 90% of AI data centre capacity sits in the United States or China. Europe's share is marginal, and its ability to scale independently is constrained by capital, energy costs and the absence of domestic chip fabrication at leading-edge nodes.

The European Commission has responded with the Chips Act, which aims to mobilise public and private investment in semiconductor manufacturing, and the Digital Decade programme, which sets targets for digital skills, infrastructure and public services by 2030. National strategies sit alongside these: France 2030, Germany's creation of a dedicated Digital Ministry, and coordinated digitalisation plans across the Nordic states. Whether these initiatives close the gap or merely acknowledge it remains uncertain. The Chips Act targets 20% of global semiconductor production in Europe by 2030, a figure that looks ambitious given the continent's current single-digit share.

Regulation: shield or drag?

Europe's most distinctive contribution to global technology governance is its regulatory framework. The AI Act, the Digital Markets Act and the Digital Services Act represent the most comprehensive attempt anywhere to impose rules on digital markets and artificial intelligence. Deloitte identifies more than 20 different national regulatory regimes across Europe that need alignment, a fragmentation that does not exist in the United States or China. The EU's approach is to harmonise these through union-level legislation, but the question running through all four scenarios is whether regulation becomes a competitive advantage or a competitive handicap.

The AI Act provides an initial framework, categorising AI systems by risk level and imposing obligations accordingly. Its supporters argue that clear rules create trust, and trust creates adoption. Its critics point out that compliance costs fall heaviest on smaller companies, the same SMEs that Europe relies on for innovation, and that the United States and China will not match Europe's standards, leaving European firms operating under heavier constraints than their competitors.

Privacy adds another dimension. Europe's stance, built on the General Data Protection Regulation, is the world's most assertive defence of personal data. Deloitte raises the possibility of a "post-privacy" world within a decade, where personal data becomes openly accessible and privacy effectively ceases to exist. Europe could resist that trajectory, but doing so may impose costs on its AI development that regions without such protections do not bear.

The venture capital shortfall

Innovation requires capital. Europe's venture capital ecosystem has stagnated or declined in recent years, according to the report, and the gap between European and American startup funding remains large. Government programmes that mobilise private capital and reduce investor risk are one lever. Corporate venture capital, which allows established companies to make strategic investments in startups, is another. The European Savings and Investments Union and the European Tech Champions Initiative are attempts to address the fragmentation of national financial markets, which prevents capital flowing efficiently to where it is needed.

The structural problem is straightforward. European savers hold large pools of capital in bank deposits and insurance products that rarely reach high-growth startups. American savers, through pension funds and public markets, channel far more into venture and growth equity. Until that allocation shifts, European startups will continue to seek American capital, and American capital will continue to extract value from European innovation.

The best case: European digital excellence

One of Deloitte's four scenarios describes a future in which Europe turns its existing strengths into global leadership. In this scenario, the continent's hidden champions, the small and medium-sized enterprises that dominate niche industrial markets, combine their domain expertise with digital innovation. The formula is summarised as "domain knowledge plus digitalisation equals global market leadership." European companies set global standards for what the report calls Industry 5.0, meaning humans working alongside advanced technologies and robotics, rather than being replaced by them.

The numbers in this scenario are striking. The European tech sector nearly doubles in size within a decade, generating annual revenues above €3 trillion, with more than half coming from digital services exports. Several former niche players grow into billion-euro global platform providers. Ten million people work in the European technology and digital economy.

The labour market transformation rests on what the report calls "dual digitalisation", drawing on the strengths of different national education systems: dual training in Germany, Austria and Switzerland; technical universities in France; practical training approaches in the Netherlands. The European single market functions as a launchpad for companies scaling globally, rather than as a series of fragmented national markets.

The uncertainties that could unwind it

Three predictions hold across all four scenarios. First, AI will be deeply embedded across all European sectors within a decade, from medical diagnostics to hyper-personalised consumer experiences. Second, cybersecurity risks will rise sharply as digitalisation expands and AI increases data demands. Third, digital-first business models will dominate, with apps and online platforms as standard customer touchpoints.

The uncertainties are equally significant. How will ethical and regulatory frameworks evolve, and will they be led by business, government or society? How will consumers and businesses respond to disruptive technologies, given that acceptance varies widely across European regions, with northern Europe generally more open and southern and eastern Europe more sceptical? Will digital platforms operate freely in Europe, or will EU regulation and American resistance create persistent tension? Will venture capital recover, or will European startups continue to migrate to better-funded markets?

None of these questions has a predetermined answer. The scenario exercise exists precisely because the range of plausible outcomes is wide.